Back to PortfolioHR Tech

AI-First HR & Workforce Management Platform

Predicts employee flight risk 6 weeks before resignation — enough time to actually retain them

🇺🇸 US🇬🇧 UK60% faster CV screening · 6-week attrition warning

Overview

Full-cycle HR and workforce management platform covering recruitment pipelines, onboarding automation, performance review cycles, and payroll processing — with AI embedded into every layer. Built for two clients: a 400-person US tech company and a 200-person UK professional services firm.

منصة شاملة لإدارة الموارد البشرية والقوى العاملة تغطي خطوط التوظيف وأتمتة التأهيل ودورات مراجعة الأداء ومعالجة الرواتب — مع الذكاء الاصطناعي مضمناً في كل طبقة.

The Challenge

Both clients were drowning in manual HR processes. CV screening took 3 weeks per role. Performance reviews were subjective and inconsistent. Most critically — both had experienced unexpected senior departures that cost them months of institutional knowledge. They needed to see it coming.

كان كلا العميلين غارقَين في عمليات الموارد البشرية اليدوية. استغرق فحص السير الذاتية 3 أسابيع لكل دور. كانت مراجعات الأداء ذاتية وغير متسقة.

What We Built

A unified platform with five modules: AI recruitment (multi-board posting, CV scoring with bias reduction, auto-generated interview scorecards), onboarding (task automation, equipment requests, buddy matching), performance (structured check-ins, AI-generated review summaries), payroll (integrated with existing systems), and retention intelligence (the AI layer).

منصة موحدة بخمسة وحدات: التوظيف بالذكاء الاصطناعي، والتأهيل، والأداء، والرواتب، وذكاء الاحتفاظ بالموظفين.

Tech Stack

ReactDjangoPythonPostgreSQLGPT-4AWS

Key Outcome

60% faster CV screening · 6-week attrition warning

Start your project →
Here's what shipped
HR Command Center
01

HR Command Center

A single morning briefing across headcount, hiring, and retention — built so an HR leader can see the whole organization's health in one glance.

  • 4 KPI cards: total employees, open positions, at-risk employees, avg time to hire
  • Per-department team health grid with status badges
  • Flight risk alerts surfaced directly on the home screen, not buried in a report
  • One-click "Schedule Conversation" from any at-risk alert
Flight Risk Prediction
02

Flight Risk Prediction

Explains why someone might leave, not just that they might — six weeks of lead time instead of finding out at the resignation letter.

  • Behavioral signal breakdown: meeting participation, collaboration, internal mobility, review sentiment
  • Flight risk score with a clear High/Medium/Low band
  • AI-generated recommended actions, ranked by priority
  • Privacy note reinforcing metadata-only analysis — no message content read
AI CV Screening
03

AI CV Screening

Ranked, bias-checked shortlists per role — screening time cut from 3 weeks to days.

  • Active postings list with applicant counts per role
  • AI match scoring per candidate with years of experience and stack
  • Bias check badge confirming no demographic filtering detected
  • Compare Top 5 for side-by-side shortlist review
Performance Management
04

Performance Management

Structured review cycles plus AI-generated team insights — goal tracking and 360 feedback in the same place as the review itself.

  • Review cycle progress: completed, in progress, not started
  • Per-employee AI review summaries, one click away
  • Team-wide AI insights flagging velocity and burnout signals
  • Bulk actions: send reminders, generate reports, export to PDF
Onboarding Pipeline
05

Onboarding Pipeline

New hires tracked from pre-boarding through month one, with smart buddy matching instead of a generic checklist.

  • Pipeline view across Pre-boarding, Day 1, Week 1, and Month 1
  • Per-hire progress bar with the current outstanding task
  • Smart buddy matching scored by shared interests and tech stack
  • Standard onboarding checklist template, customizable per department

The AI Layer

The flight risk model is the most strategically significant component. It was trained on anonymised attrition data across role types, tenure bands, team structures, and performance trajectories. It analyses 14 behavioural signals — including changes in meeting participation, performance review sentiment drift, peer recognition patterns, and manager interaction frequency — to produce a weekly flight risk score per employee. When a score crosses the threshold, HR receives a specific, actionable alert (not just "John is at risk" but "John has shown declining engagement in cross-functional meetings over 6 weeks, combined with no promotion movement in 14 months — recommend a career conversation this week"). The CV screening model was fine-tuned on role-specific competency frameworks, reducing screening time from 3 weeks to 4 days while improving shortlist quality scores by 38%.

نموذج مخاطر المغادرة هو المكون الأكثر أهمية استراتيجياً. تم تدريبه على بيانات استنزاف مجهولة الهوية عبر أنواع الأدوار وفترات الخدمة وهياكل الفرق. يحلل 14 إشارة سلوكية لإنتاج درجة مخاطر مغادرة أسبوعية لكل موظف.

Results

  • CV screening time reduced from 3 weeks to 4 days (60% faster)
  • Flight risk detected average 42 days before resignation
  • Shortlist quality improved 38% per hiring manager ratings
  • Performance review completion rate increased from 67% to 94%
  • Deployed for 3 enterprise clients across US and UK
- تقليص وقت فحص السير الذاتية من **3 أسابيع** إلى **4 أيام** - رصد مخاطر المغادرة بمعدل **42 يوماً** قبل الاستقالة - تحسن جودة القوائم المختصرة بنسبة **38%**